Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
Thanks for the advice, Andy.
I'll try this out painting a new... Roughness Map. 
I ran some tests today in DS 4.8 and DS 4.9. Can't say I'm very impressed. About Roughness (PBR Metallicity/Roughness) and Glossiness (PBR Specular/Glossiness) aka "microsurface properties". The IrayUber documentation states, that "Glossiness is the inverse of Glossy Roughness". Means, a surface that is 100% glossy is 0% rough, and one which is 100% rough is 0% glossy. On both extremes both mixes produce an almost same result. But there's a huge difference for a surface which is at 50% roughness/glossiness, when the linear values of the mixes meet half way. Both should show a similar result. But they don't. I compared my results to the Marmoset microsurface cluesheet:
There, glossy plastic has a linear value for microsurface of 0.46. Roughness at 0.0 and Glossiness at 1.0 indicate that this graphic represent the Specular/Glossiness workflow.
On my renders, the left sphere is on Metallicity/Roughness, the right one on Specular/Glossiness. Default values with Glossy Layered Weight at 1.0, Sun/Sky only, Date: 10.08.2015, Time: 13.30.00. Specular Color value on the Specular/Glossiness Mix are set to 0.40, 0.40, 0.40 (the calculated correct value for a Refraction Index of 1.50, using the Fresnel equation. Although the previev shows a 0.40 there, it is rounded. The actual value set there after converting from 3Delight to Iray is 0.0399472.)
While there should be a reflection at the "50" render on the Spec/Gloss similar to that on the Metal/Rough... it isn't, just a faint "blur". I immediately checked at my fridge, if it has taken a break and looking for a snack... no reflection there.
Bottom line: Specular/Glossines Mix in DS is broke. Very. FUBAR. Even if we had some valid BRDF measurements, for human skin, vulcan skin, plastic, whatever... they wouldn't work in there.
My guess is, that the "Reflectivity" parameter value you reclaim on the Metal/Rough Mix, is on Spec/Gloss either absent or not working as intended.
With the 4.9.0.44 Beta the results are identical. Except, that for the same dull scene DS 4.9 needs almost twice the rendertime to reach 95% convergence (the numbers (0 - 100) are the values set for "Glossy Roughness"/"Glossiness" in % [in linear space: 0.00 / 1.00]):
DS 4.8:
4.8 Metallicity vs. Specular 0 - 100
Total Rendering Time: 1 minutes 50.56 seconds
Loaded image r.png
CUDA device 0 (GeForce GT 730):
144 iterations, 4.712s init, 104.116s render
4.8 Metallicity vs. Specular 50:
Total Rendering Time: 3 minutes 16.11 seconds
CUDA device 0 (GeForce GT 730):
257 iterations, 8.924s init, 185.320s render
4.8 Metallicity vs. Specular 100 - 0:
Total Rendering Time: 2 minutes 50.73 seconds
CUDA device 0 (GeForce GT 730):
225 iterations, 7.639s init, 161.248s render
DS 4.9:
4.9 Metallicity vs. Specular 0 - 100
Total Rendering Time: 3 minutes 7.48 seconds
CUDA device 0 (GeForce GT 730):
236 iterations, 11.019s init, 174.056s render
4.9 Metallicity vs. Specular 50
Total Rendering Time: 7 minutes 5.55 seconds
CUDA device 0 (GeForce GT 730):
536 iterations, 21.004s init, 402.718s render
4.9 Metallicity vs. Specular 100 - 0
Total Rendering Time: 4 minutes 58.15 seconds
CUDA device 0 (GeForce GT 730):
378 iterations, 15.503s init, 280.798s render
16k bump maps sound terrible. That's 805 Megabytes Video RAM eaten up per bump map. Genesis 2 has three, Genesis 3 has four of them: that's 2.4 / 3.2 Gigabytes VRAM just for bumpmaps alone. 4k diffuse textures add 150/200 MB, same for translucency maps which aren't the same as the diffuse ones, another 100 MB for a decent 8k x 4k HDRi, if your card runs the monitor also will use up another around 200 MB... On a 4 GB minimum recommended graphics device there's not much RAM left for textured scenery... or even a second character in the scene (except a Silver Surfer
).
BTW, I also ran an investigation for mesh RAM usage, an untextured Genesis 2 Figure eats around 25 Megabytes VRAM. And it seems that Iray eats around 870 Megabytes by default (rendering a completely empty scene, just a camera and a light).
@Arnold C.
--- the improvment from 4k to 6 - 8k.. is allready a big gain.. (so for best possible portrait closeups - 6- 8k recommended). - you wont have 2 figures or more in a portrait closeup - 
i think you are on the right track - i always said that for a real looking skin i need a glossiness of 0.8 - 0.9 in PBR glossy.. (to get the right looking sharper highlights on a nose).... this was and is strange and somehwere in the range of polished steel but not skin... but maybe the scale is NOT linear?
Just guessing: in a PBR engine there is no light beam control (cheating)...just surfaces and material... so the parameters glossines or reflectivityy simulate surfaces.. and reflections are NOT proportional/linear to the size of the "corns" from a rough surface....this is not linear in nature... maybe you ask over at the nvidia forum what the parameter actually controls.
well - 16k is the resolution which you need to have the same details on the torso as on the face - dont worry in 2-4 years graphic cards with 12 GB are entry cards
Rashad,
Thanks for the example of "fireflies". I've benefited from your example pic but am wondering why you didn't circle the middle of the right eyebrow. Did you just not bother or is that a different problem... certainly unwanted IMO?
Thanks for prompting me to get DS 4.9, I'd been putting if off. It is better than 4.8 in many ways.
WOW.. so this is an issue others are dealing with as best they can too. Interesting. Thanks for the update to the post with that info about 4.9 beta!
As I understand it "Glossiness is the inverse of Glossy Roughness" is ment just for the values. Meaning, with glossiness value 1 you will get clear reflections, which is inverted from glossy roughness where you need value of 0 for clear reflections. It does not mean that glossiness and glossy roughness do the same thing, just inverted.
In specular/glossiness mode reflections are automatic (specular/glossy = reflective) which is why we dont have glossy reflectivity. And glossiness is reflection blur, so 0.5 glossiness (50% blur?) is one very "spread out" reflection.
well @ben98120000 i remember that one of my first questions in this forum about iray was: Does somebody know the difference between reflectivity and glossiness? for me (physically it is one and the same
.....
.
i use PBR Specular for skin because i had more problems with to much ambient reflections in PBR metal... and arnolds test showing me now the same.... one and the same glossiness map gives a complete other effect in this two shaders. And it would be great to be able to understand WHY
Iray has somehow to many glossines parameters for ONE physically simulated model (in PBR metal).. so - are there different models? Does PBR metalicity glossiness reflectivity cheat (aka.. just blur the highlights linear as a photoshop filter does? )...
Until yet it was my understanding that there are only two main models which describe reflectivity.. IOR/specular (measured light) and roughness/scattering (microsurfaces)... so why is there a parameter called glossiness reflectivity?
Here is some interesting read about the matter at hand I had hidden in my bookmarks:
http://www.marmoset.co/toolbag/learn/pbr-theory
Glossiness/Glossy Roughness are (or at least should be on our IrayUber) the two sides of the same medal named "Microsurface", sometimes mentioned as "Microfacet Theory" (surface irregularities at microscopic level causing light diffusion). "Reflection Blur" not caused by surface irregularities, but by absorption and scattering effects, is a complete different pair of shoes, and tied to the materials' Reflectivity.
The most significant difference between the PBR Metallicity and PBR Specular mix on the IrayUber is, IMO, that PBR Metallicity's "Glossy Specular" color is calculated automatically and internally (Fresnel equation), while on the PBR Specular it has to be set manually. And calculated by the user himself.
If you're assuming correctly, that reflectivity is "on automatic" in DAZ's IrayUber, then we've found our perpetrator. Did another test render, same scene, Glossy Roughness set to 0.10 on the left, and Glossiness to 0.90 on the right plastic ball (see picture below). And again comparing it to a Marmoset cluesheet, this time "Reflectivity":
Although they use a different rendering engine, the good thing about Physically Based Rendering is, that if the renderer is set up correctly, results will be (or should be) identical. Because the values don't differ whatever renderer you use. If you take a closer look at the three rightmost examples, representing a reflectivity value of 0.80, 0.90 and 1.00, there's not much of a significant difference in "blurriness" on the reflections.
Andy, I'd suggest you should read through this article (yes, Marmoset again, but IMO easier to understand than Allegorithmic's PBR Guides. And better structured). The Reflectivity topic is covered from Diffusion & Reflection to Fresnel, and the topic Glossiness/Roughness from Microsurface on.
As I understand, Reflectivity in PBS covers light diffusion on a surface due to absorption and scattering effects. And thus, yes, it does cheat. I assume NVIDIA developed this parameter to mimic absorption and scattering effects without the need to use the Subsurface Scattering capabilities of the "Volume" material struct.
Glossiness/Glossy Roughness mimic Microsurface capabilities. Each one having a different starting point on the linear space value ladder.
As you mentioned having problems to determine reflectivity on the PBR Metal... QUOTE: "Conductors will usually exhibit reflectivities as high as 60-90%, whereas insulators are generally much lower, in the 0-20% range. These high reflectivities prevent most light from reaching the interior and scattering, giving metals a very “shiny” look." END QUOTE.
This makes me wonder if the guy(s)/gal(s) at DAZ are using a correct value (0.50 by default) for an insulator like plastic is one. Judging by the use of a whacko 1.00 RGB value on Generation 7's Transmitted Color (red channel) and the PBR Specular/Glossiness discrepancies I'd bet the chance is high on this one. I also argue they already know that something in PBR Specular doesn't work well... since they always use the PBR Metal Mix for their Gen7 shaders. Which is odd if PBR Specular really would be a that better method as some people claim.
EDITH: Added another test render. In the range from 0.90 to 0.99 the blurred dot on the PBR Specular decreases in size and blurriness for each 0.01 decimal added to the 0.9 value. At 0.99 the results seem to match. Although the "Specular" plastic ball appears to be slightly darker than the "Metallicity".
reflectivity does not change the diffuse light amount with a constant albedo or IOR - but it does so in PBR metal....if something is broken or wrong named then it is This parameter
...
.. it acts just like it does more then just changing reflectivity
Had to edit.. because i was simple plain wrong here... but something with this parameter is fishy
Depends on what value (greyscale color) your Glossy Roughness was set. On my first posted roughness test render I used a value of 0.48 (0.4784313725490196 exact) for the darkest part of the T-Zone (nose), equals a RGB 122 monochrome. As you may already found out yourself, you can deternine the corresponding monochromes for a Gloss/Roughness texture map by multiplying 255 by the linear space value (0.0 to 1.0).
Both Glossy Roughness and Glossiness now range from 0.0 to 1.0, but they're answering a different question asked:
Glossy Roughness = Am I rough? (or How rough am I?)
The answer for a 0.0 value (RGB 0): you're not rough, you're glossy. And for 1.0 (RGB 255): you're insanely rough, you're not glossy.
Glossiness now wants to know how glossy it is. 0.0: not glossy, but rough. 1.0 insanely glossy, not rough.
The reason why the same texture map produces different results are hidden in the monochrome colors: on "Glossy Roughness" a RGB 0 black equals 100% glossiness, which on "Glossiness" it's a RGB 255 white (which on "Glossy Roughness" equals 100% roughness).
You would need to invert the Roughness map to work correctly for "Glossiness", and vice versa. But since someone has broke PBR Specular on the IrayUber, you can save your time, 'cause you won't get the appropriate results by just doing a simple inversion.
now when in pbr metal.. the parameter glossiness/reflectivity includes BOTH values allready... roughness = gloss ... reflection = light absorbation amount..
.. but something is fishy with reflectivity in PBR metal 
.
edit: i was plain wrong here
ok .. maybe we posted at the same time...
but i try to come to the point.... you can not get the same result 50% using the parameter GLOSSINESS (and) REFLECTIVITY because in PBR specular you just change GLOSS to the same level. But you would need to match GLOSS and IOR or specular... and this is not linear to PBR/metal
but you dont know how much gloss is needed for plastic /which kind of plastic surface).. or do we?
all we have is the idea that 50% is plastic in pbr metal... or do i overlook something?
I used shader mixer to see how bricks are connected.
In metallicity/roughness mode, three roughnesses (glossy, backscattering and refraction) are connected to their own "square roughness" brick, all of which are, amongst other things, connected to the same "roughness squared" on/off switch which is on and hidden by default.
In specular/glossiness mode, three roughnesses (glossy, backscattering and refraction) are connected to their own "gloss to rough" brick and theres no "connect them all" switch.
So did some testing. Metallicity/roughness mode with metallicty 1, white base color, glossy layered weight 1, glossy reflectivity 1, glossy roughness 0.3 and "roughness squared" turned off, gives same result regarding the reflection bluriness as specular/glossiness mode with same colors and weights where possible and glossiness 0.7.
before i get fireflies in my brain
...
.
PBR metal against PBR specular
roughness = gloss
gloss = roughness
reflection = specular
specular = reflection...
but also
roughness = specular... actually it replaces specular.. but glossines is also based on roughness which we dont have in PBR specular, because there is specular ..
Is this right?
but on what is then glossines based in PBR specular?..it should be roughness.. but that would change specular massively, which it does not
i need a drink
You either didn't read the article already I pointed you to or misunderstood what he says about Energy Conservation. It simply means that a material cannot reflect more light than it originally receives. Please don't crop cluesheets, you did left out the leftmost balls on the reflectivity one:
You can clearly see the lesser amount of diffusion (blur) the higher the reflectivity value is. And a definite no to "diffusion can only be achieved by microsurface". It's also caused by subsurface scattering effects. Reflectivity determines the amount of light reflected vs. the amount of light absorbed and scattered. The higher the reflectivity, the more light will be reflected and the less will be absorbed/scattered. You're right, reflectivity includes a kind of "roughness", but one which isn't caused by microsurface/microfacet surface material properties. Like explained by this picture:
There is no "other PBR terminology". The principles of Physically Based Rendering, or Physically Based Shading, are the same, independent of the rendering engine used and the methods of how it calculates the outcome. That's the great thing about it. In contrary to 3Delight and Posers Firefly renderer, PBR renderers are all using the same principles and measured values, even the same PBR texture maps, and all will create similar results. Where one don't, someone has messed up.
Yeah, we did.
Of course I do. And you, too. This page, first post, on the graphic... "plastic (glossy)". Has a "Glossines" value of 0.46 using a Specular/Glossiness workflow, but unfortunately is a result which you can't achieve in DAZ's PBR Specular/Glossiness Mix. At least not with this value.
Now I guess you're confusing the 0.50 for Glossy Reflectivity with the Glossy Roughness parameter of 0.32. Which should equal to Glossiness 0.68 (the inverse) but it doesn't... the damn thing.
read my postings before you press enter !.
. i allready edited my post.. saying i was plain wrong 
(partial) 
...
in a specular workflow - which does INCLUDE roughness and other parameters in specular allready.. what is glossiness?
In the DAZ3D docs they do say PBR Spec is harder to set up than PBR Metalicity/Roughness and that is way I only use PBR Metalicity/Roughness and Weighted
haha - well.. it is really my firefly in my brain night..(or to much wine)... sorry Arnold...
but my last question is actually where i think i am stuck .... i dont see how glossiness can be the oposite of roughness in a specular workflow. it can not be linear to roughness when specular includes roughness allready - what is it then?
@Szark right.. that's how i understood it too.... metal.. reflectivity is the easy to use shader for everybody... and that's maybe because specular glossines is not linear to the roughness workflow...
ben98120000 posted that the roughness bricks have other connections... so something IS here different..
I have been watching a lot of Substance Painter tutorials over the last few days and most of the SP team say that PBR Roughness model is the best to use.
The forum doesn't tell if someone edits his/her posting... it even doesn't tell if somone post a new posting even when "Bookmark" is active.
But it always tells you if something's new going on in the shop. Odd...
>>Then don't edit your postings when I'm already working on an answer!<<
The Specular/Glossiness workflow doesn't natively including roughness. Nor does roughness equals specular. Specular is "just" the amount and coloring of the specular reflection. In the PBR Metal mix the "Glossy Specular" color is calculated by the Refraction Index value set. The renderer does it internally, "under the hood". That's why we don't have a user parameter there. Choose your Refraction Index value and Iray does the rest.
In the PBR Specular Mix, the Refraction Index parameter is there, but it doesn't work. You can try and set it to 5.0, 50.0, even 100.0... your specular reflection won't change a bit. To determine specular reflection you have to use the "Glossy Specular" parameter. The user has to calculate the color value himself and put that in. [((Refraction Index value - 1) x (Refraction Index value - 1)) / ((Refraction Index value + 1) x (Refraction Index value + 1)). The result is a decimal which you have to put in in each of the RGB linear color fields of the "Glossy Specular" parameter. As an example, for a Refraction Index value of 1.50 you'll get 0.04].
"Glossiness" and "Glossy Roughness":
The Unreal 4 game engine uses a PBR renderer on the PBR Metallicity/Roughness workflow, so I borrowed their comparision chart. On the top you have the the color gradient, ranging from RBG 0 Black to RGB 255 white, with lots of different greys in between , and below them are the example renderers, for conductors (metal) and insulators (everything not a metal). As you can see, the rightmost render example with 0.0 Glossiness is considered "glossy". And the leftmost with 1.0 roughness is considered "rough". This is the Metallicity/Roughness workflow.
Now, to change from Metallicity/Roughness to Specular/Glossiness, you simply would have to flip the example renders, the color gradient stays as it is. You still have different monochrome colors determine each glossiness and roughness, depending on the workflow... except at one point: the axis where you flip at. At that point, somewhere at 0.5, the reflections of either workflow should match, no matter what. As you see on the plastic (glossy) example from the Marmoset Toolset with 0.46 Glossiness and a nice shine, they did their Specular/Glossiness set right.
That's the inversion, or the opposite. Materials are either glossy or rough, smooth or not that smooth. With a lot of stages inbetween.
Or, much simpler: a material that is 70% glossy is 30% rough (PBR Specular), and one that's 30% rough is 70% glossy (PBR Metal). The results should be identical, regardless on what end of the scale you start to look from.
this ball has ZERO reflectivity and zero roughness in PBR metal. with glossy weight 1.
therr should be NOTHING to see from the spot and fresnel reflectivity (ambient reflections on the edges) looks to me like it is fully there...
and that's why i said.. if something is broken then it is reflectivity in PBR metal. .
this example has glossy weight also on zero..
so zero weight, zero roughness, zero gloss..
fresnel is gone (that's why i have always to much ambient reflections..... just by weigthing in the glossy layer...
BUT the spot reflection is still there and looks like it is as reflective as plastic or more even with zero
and that's why i said earlier in PBR metal i need 4 parameters to control reflections
aka: roughness, gloss, weight and bump... to reach the same effect as i do in pbr specular with 1 (2) parameters...
Now we can point on formulas as long as we want (i appreciate what you do - will study it tomorrow again
And that's why convertion over to PBR specular can not be linear (IN IRAY)-- ether because PBR metal is broke.. or they use a complete other model (which include WEIGHT?)... then all other pbr engines do.
That's indeed odd... looks like glossy reflectivity and glossy roughness are independent from each other. After seeing what you did I ran a test, too. Albedo RGB 0 black, zero reflectivity, full roughness and layered weight, Refraction Index set to 1.0. 15,000 lumen spotlight at 5k. I can't see any reflections on mine. Actually, I can't see a... anything. My new family crest: matte black marble on black ground.
Changed the albedo to RGB 0, 0, 160 blue. All I can see is the albedo then. Did you point a mirror at the back of the marble?
??? Not being linear would be very bad, for both workflows. AFAIS PBR Metal is linear. Not so sure about PBR Specular. At least Glossiness seems to be more than a lil' bit off.
hmm... i must do the same again tomorrow with a clear brain
.. but based on this strange behavior (weight = full ambient reflectivity ) i dont trust the ubershader . no mirror.. hdri plus spot... my skinnmodels are all based on a weight 1.. that's why this ambient reflections are always a problem.. also in specular but lesser...
..
i see them way better because i work with a higher glossiness or lower roughness amount then usally. glossines maximum values. based on the nose which looks like polished plastic and not the average skin values which must be token on the dry back somewhere
ok... today is another day.. and no fireflies stop me from formulating what i figured out while drinking wine
...
automatic fresnel ambient reflections: this behavior is a limitation to me in both shaders aka metalicity and specular
1. the automatic ambient reflections amount is dependent only from glossiness weight
. it is not really a problem when you use low glossiness/high roughness.. but models which use mesh/map microdetails to simulate glossiness/roughness - wont react like they should.
PBR metalicity - or my slight antipathy to map roughness and use reflectivity there
2. reflectivity is to high because even set to zero.. the spot is still there... and there should be no lightbeam reflections also on a full polished marble.. zero reflection = no light bouncing back...this problem applies only to PBR metal (aka pbr metal is "broke"),
to replicate the marmorset reflectivity example... you set roughness to 1.... (to have a dull marble).. and that's not how i understand it... even with full gloss aka zero roughness in PBR metalicity.. the marble should NOT reflect something - no direct light and also not from the environment.
this behaviour does not necessarly mean that iray uber shader is complete wrong (but not 100% correct) - because such values are out of the range from natural materials.... but very irritating because it means that the slider values are not linear from 0 - 256.. as they are in the marmorset examples.
if the reflectivity slider does not start from ZERO as i showed above (marble render) .. then 50% is not the middle to the specular colors! and you wont have the same result simple with converting over to specular... and here might be the bug which confused us both.
you did exactly the opposite as me translating a wrong reflectivity scale render (visual) in the correct specular color!:
might be a bug in iray.. OR in the ubershader
if i am right (need to check 4.9 better first) and you coming to the same conclusion - then i think it is time to fill a bug report.
I found that Gloss Weight and Reflectivity were omitted as soon as you turn on Metalicty The only control is the Roughness however when using a B&W Metalicty Map these channels do control the non metal parts denoted by the black parts of the Metalicty Weight Map. And also I found that the grayscale factor of the weight or roughness maps isn't linear meaning 128 grey isn't 0.50 roughness etc.
i just installed 4.9.0.44.... and something changed exactly with the parameter reflectivity in pbr metal... (just testing the parameter right now)
So to make sure that we talk about the same problem... we must check that we always use the same version and can still replicate the error.
that's how full reflectivity with zero roughness and a spot with disc in the standard hdri looks now: image 1..
image 2 shows zero reflectivity and zero roughness.. only the irritating fresnel reflections are still there.(glossiness weight 1)
Well to be honest I am using the beta but havent had a chance to check. Interesting.